Current higher-level automation tools that support generation, selection and implementation of decision alternatives are hardly adopted in time-critical complex domains, despite findings showing that such tools can improve performance with up to 20-50%. Low trust and need to keep the competence are the main reasons.
NextGenDST is a two-year strategic research project. The overall goal of the project is to improve performance within time-critical complex domains, such as emergency management, Air Traffic Management (ATM), maritime and train dispatching, by developing a new generation decision support tool (NextGenDST). The tool will enable better human/automation collaboration during decision-making activities by:
-enabling humans to communicate their intentions, preferences and priorities to the tool
-enabling negotiation between humans and the tool
-identifying decisions that are close to the optimal ones but that humans can control, understand and execute
-intervening only if the potential for improvement is high
-explaining the rationale behind the proposed decisions in a understandable way for humans
-presenting impacts of the proposed decisions on the global performance of the system.
For example, the tool would enable the operator in a train-dispatching centre to tell the system in which sequence he plans to dispatch the trains by tactile interaction and/or augmented reality techniques. The tool would then, only if needed, propose a better but understandable sequence by using 3D animation and present how this would affect punctuality.
In the long run, this will improve human operators' trust in the decision support tool, increase their competence (help them understand why and how to make better decisions) and improve the quality of decisions they make. The project will overcome key challenges associated with use of automation, such as loss of situation awareness and out-of-the-loop syndrome by enabling better human/automation collaboration.